linux-kernel - Re: [PATCH 1/5] MIPS: Introduce irq

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:   Wed, 7 Dec 2016 09:25:53 +0000
From:   Matt Redfearn <matt.redfearn@...tec.com>
To:     "Jason A. Donenfeld" <Jason@...c4.com>
CC:     Ralf Baechle <ralf@...ux-mips.org>, <linux-mips@...ux-mips.org>,
        Thomas Gleixner <tglx@...utronix.de>,
        Paolo Bonzini <pbonzini@...hat.com>,
        Chris Metcalf <cmetcalf@...lanox.com>,
        Petr Mladek <pmladek@...e.com>,
        LKML <linux-kernel@...r.kernel.org>,
        James Hogan <james.hogan@...tec.com>,
        Paul Burton <paul.burton@...tec.com>,
        Aaron Tomlin <atomlin@...hat.com>,
        Andrew Morton <akpm@...ux-foundation.org>
Subject: Re: [PATCH 1/5] MIPS: Introduce irq_stack

Hi Jason,


On 06/12/16 22:13, Jason A. Donenfeld wrote:
> On Fri, Dec 2, 2016 at 2:39 PM, Matt Redfearn <matt.redfearn@...tec.com> wrote:
>> +void *irq_stack[NR_CPUS];
> I'm curious why you implemented it this way rather than using
> DEFINE_PER_CPU and the related percpu helper functions.
Because in the IRQ entry point in assembler we have to look up the 
address of this CPU's IRQ stack. Doing so with a simple array can be 
done with fewer instructions than a per-cpu variable. The kernel stack 
pointer for each CPU is held in a similar array.

MIPS does not have a particularly optimized per-cpu implementation with 
the per-cpu offset being held somewhere easily accessible, so right now 
it has be looked up from the __per_cpu_offset array, and then applied to 
the per-cpu pointer. Obviously doing the first lookup is analogous to 
what I am doing, and the subsequent application to the per-cpu pointer 
would be additional instructions.

Thanks,
Matt